Abstract: Apparatus and methods related to image processing are provided. A computing device can determine a first image area of an image, such as an image captured by a camera. The computing device can determine a warping mesh for the image with a first portion of the warping mesh associated with the first image area. The computing device can determine a cost function for the warping mesh by: determining first costs associated with the first portion of the warping mesh that include costs associated with face-related transformations of the first image area to correct geometric distortions. The computing device can determine an optimized mesh based on optimizing the cost function. The computing device can modify the first image area based on the optimized mesh.

Abstract: A system and method for providing unsupervised domain adaption for spatio-temporal action localization that includes receiving video data associated with a surrounding environment of a vehicle. The system and method also include completing an action localization model to model a temporal context of actions occurring within the surrounding environment of the vehicle based on the video data and completing an action adaption model to localize individuals and their actions and to classify the actions based on the video data. The system and method further include combining losses from the action localization model and the action adaption model to complete spatio-temporal action localization of individuals and actions that occur within the surrounding environment of the vehicle.

Abstract: Embodiments provide methods and apparatuses to manage hardware resource access on a mobile device comprising detecting an attempt by the first application to access the first hardware resource when the application is executing in the background and determining that there is not a policy for the first application to access the first hardware, providing a notification requesting user input thereof, and adding a new policy accordingly.

Abstract: Photorealistic image stylization concerns transferring style of a reference photo to a content photo with the constraint that the stylized photo should remain photorealistic. Examples of styles include seasons (summer, winter, etc.), weather (sunny, rainy, foggy, etc.), lighting (daytime, nighttime, etc.). A photorealistic image stylization process includes a stylization step and a smoothing step. The stylization step transfers the style of the reference photo to the content photo. A photo style transfer neural network model receives a photorealistic content image and a photorealistic style image and generates an intermediate stylized photorealistic image that includes the content of the content image modified according to the style image. A smoothing function receives the intermediate stylized photorealistic image and pixel similarity data and generates the stylized photorealistic image, ensuring spatially consistent stylizations.

Abstract: Embodiments provide methods and apparatuses to manage hardware resource access on a mobile device comprising detecting an attempt by the first application to access the first hardware resource when the application is executing in the background and determining that there is not a policy for the first application to access the first hardware, providing a notification requesting user input thereof, and adding a new policy accordingly.

Abstract: A semiconductor device may include a substrate and spaced apart first and second doped regions in the substrate. The first doped region may be larger than the second doped region to define an asymmetric channel therebetween. The semiconductor device may further include a superlattice extending between the first and second doped regions to constrain dopant therein. The superlattice may include a plurality of stacked groups of layers, with each group of layers comprising a plurality of stacked base semiconductor monolayers defining a base semiconductor portion, and at least one non-semiconductor monolayer constrained within a crystal lattice of adjacent base semiconductor portions. A gate may overly the asymmetric channel.

Abstract: A method for making a semiconductor device may include forming spaced apart first and second doped regions in a substrate. The first doped region may be larger than the second doped region to define an asymmetric channel therebetween. The method may further include forming a superlattice extending between the first and second doped regions to constrain dopant therein. The superlattice may include a plurality of stacked groups of layers, with each group of layers comprising a plurality of stacked base semiconductor monolayers defining a base semiconductor portion, and at least one non-semiconductor monolayer constrained within a crystal lattice of adjacent base semiconductor portions. The method may also include forming a gate overlying the asymmetric channel.

Abstract: A user can issue a query about a process that has a number of stages. A stage of the process is determined using the query and location data associated with the query, and a stage prediction model. A stage learning system can select a sample of query logs for a category from a database of millions or billions of users' queries. Queries can be parsed into keywords. A category can be determined from location information associated with each query and from query keywords. Queries are aligned based on location and, optionally, keywords. TF-IDF values are computed for queries and are used to determine a difference significance between aligned, adjacent queries. If aligned, adjacent queries have a substantially difference in keywords and TF-IDF, then a conversion stage is identified. Content can be presented to the user based on the category, keywords, location, and conversion stage.

Abstract: Segmentation is the identification of separate objects within an image. An example is identification of a pedestrian passing in front of a car, where the pedestrian is a first object and the car is a second object. Superpixel segmentation is the identification of regions of pixels within an object that have similar properties. An example is identification of pixel regions having a similar color, such as different articles of clothing worn by the pedestrian and different components of the car. A pixel affinity neural network (PAN) model is trained to generate pixel affinity maps for superpixel segmentation. The pixel affinity map defines the similarity of two points in space. In an embodiment, the pixel affinity map indicates a horizontal affinity and vertical affinity for each pixel in the image. The pixel affinity map is processed to identify the superpixels.

Abstract: A spatial linear propagation network (SLPN) system learns the affinity matrix for vision tasks. An affinity matrix is a generic matrix that defines the similarity of two points in space. The SLPN system is trained for a particular computer vision task and refines an input map (i.e., affinity matrix) that indicates pixels the share a particular property (e.g., color, object, texture, shape, etc.). Inputs to the SLPN system are input data (e.g., pixel values for an image) and the input map corresponding to the input data to be propagated. The input data is processed to produce task-specific affinity values (guidance data). The task-specific affinity values are applied to values in the input map, with at least two weighted values from each column contributing to a value in the refined map data for the adjacent column.

Abstract: Segmentation is the identification of separate objects within an image. An example is identification of a pedestrian passing in front of a car, where the pedestrian is a first object and the car is a second object. Superpixel segmentation is the identification of regions of pixels within an object that have similar properties An example is identification of pixel regions having a similar color, such as different articles of clothing worn by the pedestrian and different components of the car. A pixel affinity neural network (PAN) model is trained to generate pixel affinity maps for superpixel segmentation. The pixel affinity map defines the similarity of two points in space. In an embodiment, the pixel affinity map indicates a horizontal affinity and vertical affinity for each pixel in the image. The pixel affinity map is processed to identify the superpixels.

Abstract: The present disclosure provides an electronic device that includes a display device. A first panoramic image capturing device is movably configured on the display surface of the display device. A second panoramic image capturing device is movably configured on a back panel relative to the display surface. A drive unit is configured to push the first panoramic image capturing device out of the electronic device along a first direction perpendicular to the display surface and push the second panoramic image capturing device out of the electronic device along a second direction, which is opposite to the first direction, when receiving a first signal. The drive unit is configured to pull the first panoramic image capturing device back to the electronic device along the second direction and pull the second panoramic image capturing device back to the electronic device along the first direction when receiving a second signal.

Abstract: Route segment information is received from a first autonomous vehicle over a network. The route segment information includes one or more route segments and estimated entrance times, each of the entrance times being associated with one of the route segments and representing a time the first autonomous vehicle enters the route segment. Whether the first autonomous vehicle is within a specific route segment is determined based on an entrance time associated with the specific route segment and a current location of the first autonomous vehicle. If the first autonomous vehicle is within the specific route segment, a request is sent to invite the first autonomous vehicle to join a vehicle group of one or more vehicles associated with the specific route segment. The vehicle group is one of a plurality of vehicle groups that were determined based on route segment information received from a plurality of autonomous vehicles.

Abstract: An electronic device including a display, a host, a hinge, a panoramic camera, a linkage assembly and a torsion spring is provided. The display has a plane facing the host. The hinge is connected between the display and the host. The panoramic camera is disposed in an accommodating space on an upper side of the display and is rotatable relative to the display. The linkage assembly is disposed in the display and driven by the hinge to limit the panoramic camera, such that the panoramic camera is kept in a state of being protruded from the plane of the display, and the torsion spring is disposed on a lower portion of the panoramic camera, wherein when the display is closed relative to the host, the hinge drives the linkage assembly to release the limitation on the panoramic camera, and the torsion spring enables the panoramic camera to be restored to the original position.

Abstract: The disclosure relates to determining thermal insulation levels of clothing a user should wear when traveling to a destination. To do this, a thermal insulation level of clothing worn by a user and levels worn by a crowd can be analyzed from images including the user or the crowd. A thermal sensitivity bias for the user can be determined by comparing those levels for similar locations. Environment data can then be collected for the destination. A thermal insulation level of clothing worn by the crowd for the destination can be predicted based on the environment data. This level can be adjusted with the bias for the user to generate a thermal insulation level of clothing to be worn by the user at the destination. An image having clothing with that level can be displayed to recommend a style of clothing for the user to wear at the destination.

Abstract: A machine learning method and a machine learning device are provided. The machine learning method includes: receiving an input signal and performing normalization on the input signal; transmitting the normalized input signal to a convolutional layer; and adding a sparse coding layer after the convolutional layer, wherein the sparse coding layer uses dictionary atoms to reconstruct signals on a projection of the normalized input signal passing through the convolutional layer, and the sparse coding layer receives a mini-batch input to refresh the dictionary atoms.

Abstract: An optical image capturing system and an electronic using the same are disclosed. The optical image capturing system includes at least two lenses, an image plane, and a first position element. In certain condition, the design of the optical image capturing system may simantaneously inhibit the deviation of effect focal length and improve the imagining quality in response to temperature fluctuation.

Abstract: The invention provides a key structure, which includes a carrying plate, a trigger member, a keycap, and a reciprocating member. The trigger member is disposed on the carrying plate, and includes a bearing portion, a pillar portion, a trigger portion and a connecting portion. The bearing portion and the trigger portion are respectively disposed at two opposite sides of the pillar portion, and the pillar portion is perpendicular to the carrying plate. The trigger portion protrudes toward the carrying plate. The connecting portion is circularly disposed at one side of the pillar portion facing the carrying plate, extends from the pillar portion toward the carrying plate, and is obliquely connected to the carrying plate. The keycap is disposed on the bearing portion. The reciprocating member comprises two opposite ends respectively connected to the carrying plate and the keycap.

Abstract: Responsive to sensor data received from one or more sensors of an autonomous vehicle, one or more predicted trajectories are generated, with each of the predicted trajectories having an associated probability. One or more driving scenarios that trigger gesture recognition are identified. For each of the identified driving scenarios, one or more gestures from one or more vehicles are detected in accordance with a gesture detection protocol. One or more gestures from the autonomous vehicle are emitted for communication with the vehicles in accordance with a gesture emission protocol based on the detected gestures. The predicted trajectories are modified based on the detected gestures, the emitted gestures and the associated probabilities of the predicted trajectories. The autonomous vehicle is controlled based on the modified predicted trajectories.

Abstract: A semiconductor device includes a semiconductor substrate, a plurality of isolation structures on the semiconductor substrate, and a plurality of blocking structures disposed directly over the isolation structures. The blocking structures have a lower reflectivity than the isolation structures.